The need for a reliable method for estimating dietary consumption in the past has become critical in case referent epidemiological studies of the relationship of diet to cancer. The research objective of this study is to estimate the reliability of estimates of past dietary intake which are derived from present diet histories or from retrospective diet histories. A population on which a current diet history was performed in 1960 is located in Evans County, Georgia. Under this proposal this population would be reinterviewed about their recollections of their diet in 1960, and their current diet in 1978. The reliability of the retrospective diet history will be assessed by comparing it to the diet history obtained in 1960. The reliability of the current diet history as an index of past dietary intake, as well as the degree of dietary change in this population in 18 years, will be assessed by comparing the 1978 current diet history to the 1960 history. Comparisons will be made of average consumption of specific food items. Correlation coefficients of individual consumptions as reported on the respective surveys will also be calculated. Rigorous evaluation of the available methods for assessing past diet is needed. This study will contribute towards a quantitative estimate of their value. Should their reliability prove low, waste of resources on unreliable studies will be avoided. Should their reliability prove high, studies will not fail to be undertaken because of unwarranted negative prejudice against these methods.